Introduction to ∞-Topoi and Exodromy for Schemes Oberseminar of the Prof

Introduction to ∞-Topoi and Exodromy for Schemes Oberseminar of the Prof

Introduction to 1-topoi and exodromy for schemes Oberseminar of the Prof. Schmidt’s AG ”Arithmetische Homotopietheorie” Pavel Sechin Time and place: SS19, Tuesday, 11:00-13:00, SR 3, Mathematikon The goal of this seminar is two-fold. The first part of the semester will be dedicated to introduction of 1-categories using the language of quasi-categories. The second part will cover the proof of the main theorem of [BGH-Exo] which claims that the functor from reduced normal schemes over a finitely generated field of characteristic zero which assigns to a scheme X a category Gal(X) of points of the etale´ topos of X is fully faithful to the category of topological categories over the etale´ fundamental groupoid of k 1. The idea of the proof is the possibility to reglue the etale´ 1-topos of a scheme from the etale´ 1-topoi of points (which can be reconstructed from the classifying spaces of the absolute Galois groups) using specialization maps between these points. This gluing process is performed in the setting of 1-categories and, in fact, is a corollary of a purely 1-categorical statement generalizing Hochster duality. Quasi-categories are certain simplicial sets which generalize both Kan complexes (i.e. simplicial sets which allow to model the homotopy theory of CW-complexes) and nerves of small categories (i.e. effectively the study of categories, since the nerve functor is fully faithful). The need for infinity categories stems from drawbacks of other approaches to incorporate homotopy theory in different areas of mathematics which we briefly recall. Arguably, the first such attempt is the notion of a triangulated category due to Verdier. One of the problems which one encounters here is that the category of functors between two triangulated categories in general does not possess a triangulated structure (as a consequence of the non-functoriality of the cone). This comes up when one seeks to glue triangulated categories (in general, to compute a limit of a diagram of triangulated categories), e.g. trying to obtain the derived category of sheaves on a ’space’ X from derived categories of sheaves on a closed ’subspace’ and its open complement. The problem of gluing can be solved provided that one works with DG-enhancements of triangulated categories. In fact, (pre-trinagulated) DG-categories can be identified in a precise manner with certain ’stable’ 1- categories, thus, showing that this approach is at least not worse than 1-categorical under appropriate circumstances. However, neither DG-categories nor stable 1-categories will be discussed in the course of the seminar. Another approach to abstract homotopy theory are model categories introduced by Quillen. They allow to study localizations of categories with respect to a set of morphisms in efficient manner. The category of functors between model categories has a model structure under mild restrictions on the target model category, thus, allowing to deal with gluing problem. However, a serious drawback comes from the fact that the study of the localization in this approach depends not only on the set of morphisms to localize, but on additional and non-canonical data. Thus, for example, to compare two localizations in this approach one needs also to compare this additional data which may or may be not possible depending on the choices made. Though having disadvantages, model categories remain a tool for the study of quasi-categories as well as a source of examples of 1-categories via certain nerve construction. The particular emphasis of the study of 1-categories in the seminar will be on properties of so-called 1-topoi, an analogue theory of Grothendieck topoi in the higher category theory. Given a Grothendieck topology on a category C, one can produce an 1-topos of sheaves of 1-groupoids2 on it which is closely related to the study of the homotopy category of simplicial sheaves on C due to Jardine. Broadly speaking, one can view an 1-topos as a non-abelian generalization of the derived categories of sheaves of abelian groups on a topos. As an example of a previous construction, there is a way to associate to every space (more precisely, to every Kan complex) an 1-topos. It turns out that when restricted to the 1-category of pro-π-finite spaces3 this functor is fully faithful. Thus, one characterize the topoi in the image of this functor (called Stone or profinite) via certain topological spaces. A particular example of a Stone 1-topos is the 1-topos of etale´ sheaves on a spectrum of a field which corresponds to the classifying space of the profinite absolute Galois group of the field via this construction. The setting of 1-categories provides a convenient environment to glue pro-π-finite spaces. As an example, which is used in [BGH-Exo], one can glue classifying spaces of absolute Galois groups of schematic points into what is called a stratified profinite etale´ homotopy type. It is an object of pro-π-finite spaces together with a stratification over a spectral topological space underlying the scheme. Gluing 1-topoi (and topoi), on the other hand, is not so straight-forward. Deligne has defined an oriented pullback − topos X × S Y for a pair of geometric morphism of topoi f : X ! S, g : Y ! S which satisfies a certain universal property in 2-category of topoi given by the diagram 1the functors in this category have to be continuous and preserve minimal objects 2which are often called sheaves of spaces, or just sheaves 3i.e. a pro-completion in 1-categorical sense of those spaces which have finite homotopy groups and are n-connective for some n 1 − q∗ X × S Y Y p∗ = g∗ (τ X S , f∗ where τ : g∗q∗ ! f∗p∗ is the natural transformation. Moreover, Gabber showed that if f; g are coherent morphisms of + ∗ ∗ coherent topoi, Λ is a ring, then for F 2 D (Y; Λ) the canonical morphism f Rg∗F ! Rp∗q F is an isomorphism in D+(X; Λ). The paper [BGH-Exo] performs this construction and proves this base change result for 1-topoi. This allows to glue the 1-topos of etale´ sheaves on points of X to a stratified 1-topos of etale´ 1-sheaves on X. The fully faithfull- ness result about pro-π-finite spaces above then extends to show that this stratified 1-topos can be reconstructed from the stratified profinite etale´ homotopy type. Finally, the main theorem follows from the result of Voevodsky which claims that morphisms between normal reduced schemes over a finitely generated field of characteristic zero can be reconstructed from the morphisms between corresponding etale´ topoi. Prerequisites. We will assume familiarity of the participants with simplicial sets, model categories, and, in partic- ular, the Quillen model structure on simplicial sets. In order to have a feeling of the objectives of the first part of the seminar everyone is invited to read through [LurHTT, Chapter 1] which is a discussion of goals, difficulties and some results (mostly, without proofs) about 1-categories. We will not spend much time introducing and motivating main definitions of the theory of quasi-categories from there. We will follow [LurHTT], but it seems virtually impossible to cover the whole material even of Chapters 2-6 in one semester. Thus, there bound to be some omissions. The first such omission is the proof of the Joyal extension theorem and related facts from [LurHTT, Chapter 1]. Note, however, that the proof of all these facts is obtained in [LurHTT], though, there exists another approach due to Joyal himself. Many other results will also be used without proof, however, the ambitious goal of the seminar is to cover a ’dense’ subset of relevant statements so that the participants of the seminar would be able to read and understand the proofs on their own. Schedule4 Part I. 1-categories towards 1-topoi. Our ultimate goal of the first part is to understand what an 1-topos is, and, in particular, how can one attach to a Grothendieck topology on 1-category (or 1-category, provided, we understand what a Grothendieck topology means there) an 1-topos. With this knowledge we could obtain, for example, the etale´ 1-topos Xet´ , which plays an important role in [BGH-Exo]. By definition, an 1-topos is a left exact localization of 1-category of presheaves of spaces on a 1-category. Thus, we need to define 1-categories, spaces and presheaves, as well as obtain methods of working with them first. The role of spaces in 1-categories is played by so-called 1-groupoids, which are just Kan complexes in the approach of quasi-categories. In comparison to the classical theory, we see that the role of a set is replaced by a generalization of a groupoid, which yields a new point of view on presheaves of spaces. In classical category theory functors from a category C to the category of groupoids can be identified via the so-called Grothendieck construction with a category cofibred in groupoids. Left fibrations are generalizations of this phenomenon to 1-categories, however, the equivalence between them and functors to Kan complexes is not so easy to obtain since it has to be true only up to homotopy (in a suitable sense). This equivalence is obtained using the language of model categories and, in particular, the covariant model structure constructed in Talk 2. Moreover, this result allows to compare the theory of quasi- categories and the theory of simplicial categories (using relevant model structures) in Talk 3. A more general version of the previous phenomenon comes up when one wants to study functors cofibred in 1-catgories and not just 1-groupoids (aka spaces).

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    5 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us